Network redundancy is the practice of building spare capacity into a network, such as extra connections, devices, power and routes, so that the failure of one part does not take a site, application or service offline. A redundant design pairs each critical component with a backup that can take over, either automatically or by a planned manual step. The goal is not zero failures, which no design delivers, but fewer outages that users notice and shorter ones when they happen.
At a glance
- Redundancy means having a spare; failover is the process that moves traffic onto it.
- A backup only helps if it fails independently of the primary, so shared paths, providers, equipment and power all count.
- Common building blocks are a second circuit, a cellular backup, paired routers or firewalls, and backup power.
- SD-WAN and BGP are common ways to use redundant links automatically; how fast they switch depends on the design.
- The right level of redundancy depends on what an hour of downtime costs the site.
What problem it solves
Many business networks have more single points of failure than their owners realize. A site with one internet circuit, one router and one firewall stops working when any of those fail, when a construction crew cuts the cable down the street, or when the provider has an outage in its local network. Phones, cloud applications, payment systems and remote access stop with it.
Redundancy removes the single points of failure that matter most. The hard part is usually not buying a backup but making sure the backup is genuinely independent. A common surprise after an outage is that the “backup” line ran through the same conduit, entered the building through the same room or relied on the same router as the primary, and failed at the same moment.
How it works
Redundancy is applied in layers. Each layer has its own failure modes and its own way of switching to the spare.
Connections. A second internet or WAN circuit, ideally from a different provider over a different physical route. Path diversity addresses the physical route; carrier diversity addresses the provider. Many smaller sites use broadband or cellular failover as the backup to a fiber primary.
Devices. Paired routers, firewalls or SD-WAN appliances, often in an active/standby arrangement, so a hardware failure does not cut off the site even when two circuits are in place.
Power and space. Uninterruptible power supplies and generators for network gear, and separate power feeds where the building offers them. A redundant circuit does not help if its equipment loses power with everything else.
Routing and switchover. Something has to notice the failure and move traffic. SD-WAN appliances monitor each link and can steer traffic away from a failing one, often quickly enough that many calls survive. Sites with their own IP addresses can run Border Gateway Protocol (BGP) with two providers so inbound traffic also finds the surviving link. Simpler setups rely on a router’s failover rules, which may take longer and may change the site’s public IP address.
Active/active or active/standby. In active/active designs both links carry traffic all the time, so the backup is exercised daily. In active/standby the backup sits idle until needed, which costs less on some services but hides problems until the day you need it.
For help designing redundant sites and choosing backup links, see our SD-WAN solution page.
When it matters for buyers
- After an outage. The first question is which single component failed, and whether anything else in the design would have failed the same way.
- Moving voice and applications to the cloud. Once phones, CRM and file storage live online, the internet connection becomes the site’s critical path.
- Opening a new office. Ask about a second provider and a second building entrance before the lease is signed; adding them later can cost more and take longer.
- Renewing circuits. Renewal is a natural time to replace a same-provider backup with a genuinely independent one.
- Cyber insurance and customer audits. Some questionnaires ask how critical sites stay connected during an outage.
Questions to ask vendors
- Which components in our proposed design are still single points of failure?
- Do the primary and backup circuits share any conduit, building entrance, central office or provider network?
- How is failover triggered, how long does it usually take, and do calls and sessions survive it?
- Will our public IP addresses change when traffic moves to the backup link?
- Is the edge equipment paired, and what happens if the primary router or firewall fails?
- How will we know the site is running on its backup, and how often is failover tested?
- What does the SLA commit to for each circuit, and how is a joint outage handled?
How it differs from high availability
High availability (HA) is the outcome: a system that stays up for a target share of time, often stated as a percentage. Network redundancy is one of the main ways to get there, alongside monitoring, failover automation, maintenance practices and well-run change control. A network can have redundant parts and still fall short of high availability if failover is slow, untested or manual, or if the redundant parts share a hidden dependency.